We have developed a rabbit model of aerosol M. tuberculosis (Mtb) infection that mimics human progressive active TB or latency, depending on the characteristics of the infecting bacilli. We found that a phenolic glycolipid of Mtb (PGL-tb) enables the bacilli to evade and/or subvert host protective immunity. Infection of rabbits with a clinical isolate that produces PGL-tb leads to continued bacillary replication in the lung and development of chronic cavitary disease;infection with a clinical isolate that does not produce PGL-tb leads to growth control and ultimately latent infection. Our present proposal focuses on further developing the rabbit aerosol infection model as a tool to understand mechanisms underlying TB latency and reactivation. We will further explore the role of two mycobacterial factors in regulating Mtb entry into latency and/or reactivation, (a) Using paired isogenic Mtb strains that differ in their ability to produce PGL-tb, we will examine the specific role of PGL-tb in modulating the host immune response that drives M.tb into latency. In these studies, we will investigate the nature of the immune response and the physiologic state of the bacilli in tissues from infected rabbits during the development of strain-dependent latency versus chronic progressive granulomatous disease (Specific Aim 1). (b) Reactivation of stationary Mtb to a replicating state in vitro has been associated with production of a small secreted protein, the resuscitation-promoting factor (Rpf). Our recent mouse studies suggest that deletion of selected rpf genes results in the failure of Mtb to persist in the lung. Using Mtb strains lacking 4 of the 5 rpf genes, we will assess the role of Rpf in establishment of latency and reactivation of latent disease in infected rabbits (Specific Aim 2). To achieve our aims, we will continue to develop and test new assays and reagents for use in the rabbit. These studies should provide insight into the molecular basis of latency and progressive disease and are likely to advance the field by validation of a much needed additional animal model for TB research. Roughly one third of the world's population is estimated to be latently infected with M. tuberculosis. Development of active disease versus latency and reactivation of latent Mtb infection appear to involve complex interactions between the pathogen and the host immune response that are not fully understood. To understand these interactions, we have developed a rabbit model of aerosol Mtb infection that mimics human progressive active TB or latency. We will further characterize this unique model and use it to examine some underlying mechanisms that drive Mtb infection towards latency or reactivation disease. Results of our studies will contribute to improved global TB control.